NEW YORK (GenomeWeb News) – California researchers reported online last night in PLoS Computational Biology that they have used existing gene expression data to find a trio of protein biomarkers in the blood for organ transplant rejection.
"We came up with this idea to let the RNA changes in the tissue tell us which proteins to go look for in the blood," co-senior author Atul Butte, a biomedical informatics and pediatrics researcher at Stanford University, told GenomeWeb Daily News, explaining that the new gene expression-based proteomics method is intended to overcome some of the sensitivity limitations of existing proteomic approaches.
The Stanford University and Lucile Packard Children's Hospital research team scoured public messenger RNA data to find mRNA changes corresponding to organ rejection. Using this approach, they tracked down three proteins found at higher levels in the blood of those experiencing acute organ rejection.
Based on their findings for acute transplant rejection — combined with preliminary analyses of other conditions — the researchers argue that the same approach provides a promising avenue for finding blood and/or urine biomarkers for other diseases as well.
The current method of testing for acute transplant rejection involves functional monitoring of the transplanted organ, and, if necessary, biopsies of some organ tissue to test for damage the researchers noted.
"It's a pretty invasive procedure to even test for rejection and a lot of people have been looking for blood-based tests to do this," Butte explained.
Although past studies have uncovered potential biomarkers associated with specific types of organ transplant rejection, he added, researchers have not yet found common markers that can be used to look for transplant rejection across organ types.
In particular, Butte and his colleagues were interested in finding blood biomarkers based on protein rather than RNA levels. "What we wanted to do is we wanted to find a blood-based protein for this, because any protein test really trumps any RNA test — it's just that much easier to do protein [enzyme-linked immunosorbent assays] than quantitative PCR, for example, for RNA," he said.
To explore this possibility, lead authors Rong Chen, a Stanford software developer, and Tara Sigdel, a biostatistician at Stanford, and their co-workers first compiled data from the publicly available Gene Expression Omnibus database using a method called integrated RNA data driven proteomics to see if they could use mRNA expression changes to predict which proteins may serve as biomarkers for dozens of diseases.
After showing that they could find known biomarkers for more than half of the diseases tested, the researchers turned their attention to transplant rejection, using the IRDDP method to assess microarray data from three studies involving 39 child and adult kidney transplant recipients and 63 adult heart transplant recipients.
The researchers initially identified 45 shared genes that were expressed at higher levels in transplant recipients experiencing organ rejection in all three studies.
From there, they focused on 10 candidate proteins that could be detected with existing ELISA kits, evaluating the levels of these proteins in 19 individuals experiencing acute transplant rejection and 20 controls.
Indeed, the team found that three of the proteins — the pro-inflammatory proteins CXCL9, CD44, and PECAM1 — were present at higher levels in the blood of kidney transplant recipients with acute organ rejection. Their follow-up work revealed that the same proteins were also found at higher levels in the heart transplant rejection group.
While not all inflammatory proteins are found at higher levels in the blood of those experiencing transplant rejection, Butte explained, the new biomarkers — along with several of the 35 potential biomarkers yet to be tested — seem to belong to a specific inflammatory pathway.
"Inflammation is a pretty broad set of processes," he said, "but I think we've got it narrowed down to one core process involved with acute rejection across these organs."
When they looked specifically at the PECAM1 protein levels in the kidney transplant recipients, the researchers found that the biomarker could detect acute rejection with 89 percent sensitivity and 75 percent specificity. Meanwhile, the CXCL9 marker had 78 percent sensitivity and 80 percent specificity, while the sensitivity and specificity for CD44 were 80 percent and 75 percent, respectively.
Based on their findings so far, the team plans to develop an ELISA-based test for acute organ rejection using one or more of the biomarkers. Their goal is reportedly to have a commercially available test within the next three to five years.
To that end, Butte said the researchers have filed for patents based on the current study but that these have not been licensed so far.
The researchers are in the process of verifying their findings using data on individuals who have received liver and lung transplants and looking at the possibility of doing prospective studies aimed at meeting US Food and Drug Administration guidelines for diagnostic biomarker tests, Butte noted.
The team also plans to use a similar strategy to look for biomarkers that can be used to predict and/or diagnose other diseases and conditions, including pancreatic, ovarian, breast, and prostate cancer.
"Bioinformatics to me now is much more than just building computational tools and websites — bioinformaticians can really be scientists. We can see a medical need, we can start with the public data, and really turn this crank and come up with diagnostics," Butte said. "It's really about translating all that data into discoveries."